Tải bản đầy đủ (.pdf) (19 trang)

Lý thuyết cực trị và ứng dụng trong đo lường rủi ro tài chính

Bạn đang xem bản rút gọn của tài liệu. Xem và tải ngay bản đầy đủ của tài liệu tại đây (503.22 KB, 19 trang )

 t cc tr  ng d  
ng ri ro 

c Th

ng i hc Khoa hc T 
Luc
: 60 46 15
ng dn: TS.Trn Tr
o v: 2011


Abstract: 
(1943) v c trm v min hp dn ci,
 Q--c tr.
ng dt cc tr ng r
ng ri ro.

Keywords: ng; R   ;  t; Th  
hc

Content
Lung:
Chng 1: Tp trung l r c kt qu ch ca l thuyt cc tr. 
  qu cho phn ng dng chng 2. i cc tr, xp x
i vt ngc tr cho khi cc u kin c
 mm trong min hp dn ci G.
1 . Chi ting Fisher and Tippet (1928), Gnedenko (1943))
c tr.
2 Min hp dn ca mt phn phi G.
3 H phn phi vt ngng


4 Phn phi Pareto tng qu
5 H phn v
6 Biu  QQ-PP
7 c lng c m h cc tr
8 Mt s m h cc tr m rng v mi lien h c m hh
Chng 2:  dng l thuyt cc tr trong o lng ri ro t ch.
1. Ri ro t ch
2. M h o lng ri ro
3. Tham s ho bin li nhun v bin thua l
4. Mt s phng ph t c  ri ro
5. Phng ph t gi tr ri u t vn
6. ng dng l thuyt cc tr trong m h ho ui chui li sut chng
kho
7.  dng  o lng ri u t c phiu ACB.
I. C kt qu ch ca lun vn t c:
II.
 ng r

 i ro cht ch
2.2 ng ri ro
Trong qun tr ri nu ch n thun d
 ng h ng
 l c rn thi ro thc ch h
 c chn ca kt qu  i ta thng s d      
lng ru ch ng r
Giá trị rủi ro (Value at Risk-c s d
bin trong qun tr ri ro th trng ca danh mc. My, VaR vng hn
ch  phng dit ln thc tit s   dng
 Tổn thất kỳ vọng (Expected Shortfall-ng ri ro th trng.
2.2.2  i ro cht ch

2.2
u t c t chc) nm gi mt danh mc. t i
m hin ti, (t+1) m cui ca k u t (thm trong tng lai), V
t
, V
t+1

  ca danh mc t

c th

m t, t+1 tng  V
t
t, V
t+1
cha
bin ngm gi danh mu t s i mt vi ri ro:
u t s b thua l, tn tht nu V
t+1
< V
t
c thua l: X = V
t+1
- V
t
n
ngV 
    t th       i ro), mt ch  nh
lng va th hin m ri ro ca danh mc (mc thua l)  bt k ngun gc
ng ca th trng, t t, v n va thun tin cho

n tr?
  i ro cn phng nhu c b 
hc tin?
a nha th k trt s 
gi [10u vn   xut mt mô hình lý thuyết về độ đo
rủi ro c g ri ro cht ch ng ri ro ca danh mc.

2.2 i ro cht ch
Hong ca th trng bng
 t
( , , )P
. Gi X
0
n ngu
u hn (hu nh chc chX  X
0
i.
u t tham gia th trc nm gi danh mc. R 
ca vic nm gi danh mc biu hin bi mc thua l tim n sau k u t 
i bin ngXX.




2.1: 

: X 

gĐộ đo rủi ro ca danh mc. Danh mc vi
mc thua l X c ri ro


(X). Trong qun tr xem

(X) nh
khon d n th chp,
 i ro

(X) gĐộ đo rủi ro chặt chẽ nu thu ki)
sau:
 T1: Bt binh tin:
Vi mi XX, a

:

(X+a) =

(Xa
 T2: C di:
Vi mi X
1
, X
2
X  :

(X
1
+X
2
) 


(X
1
) +

(X
2
)
 T3: Thun nht dng:
Vi mi XX

X

(X)
 n 
Vi X
1
, X
2
X X
1
 X
2
(hu nh chc ch

(X
1
) 

(X
2

)
 gii t nh sau:
 T1: Vi danh m ri ro

(X), khi b n phi r a 
m ri ro ca danh mc gi

(Xa.
 T2: Ri ro ca danh mc tng hp (ng vi X
1
+ X
2
n ln tng ri ro ca
anh mp v Đa dạng hóa đầu
tư.
 T3: Danh mn.
 T4: Danh mc thua l tim 
Nh vy tt c i v u h p vi
thc tin.
 i ro ca danh mp ct t
n tr r  nguồn gốc của rủi ro   th.
Kt qu sau s cung c  i ro c th  
ri ro cht ch 

2.2u di

i ro cht ch
(pu gi
1
0

( ) 1x dx



.
 xem (p) nh mt d t.
Gi F(xi ca mc tn tht X i ro nh sau:
1
1
0
( ) ( ) ( )X F p p dp




(2.1)
vi (p(p
1
0
( ) 1x dx



 coi (p trng.







(xem [10]):
 t ch  khi (p)A.
Vc biu din (2.1) c i ro cht ch  
 th p vi ngun gc ch(p
gt quan trng ri ro.


2.3  ri ro (Value At Risk)
2.3.1 Ngun gn
Thut ng  ri ro (Value at Risk  

c s dng r
c s tr t



m quan trng trong khoa hc kinh t t sau s kin th
trng ch 
Ngp c 
 la chn danh mu t 
trn hip phng sai li su n phi u danh mu t.
Trong nhu thy ban Chi h
K (SEC -   bu
cu v vn l   tin cy 95% trong
khong th  i li su c s dng
 n l ting thn
khai s dng mt h th vu t c
Trong thi gian cui thu tht s t chc hin
 h tr cho vi vu t n ch ri ro ca th trng.
Nhng s kiu nhy rt nhip

rc rn th di mc d kin honh m
 trc. T khi tt c n vi tn tht,
 u kin tt y ri ro ca hu h
ty. Ti ting - Dennis Weatherstone 
 i tng kt tn tht ca tt c 
 tra giao d 
tn thn cho mc ng dng rt
t rt nhiu ph
p mi li c lng cn thi
l gii hn trong m 
t
c lt phn c
     t rng tt c     t ph  
nh lng v hoa h. Nh
 bm c  
ca h.
T chc t  u t 
n nh y hn na vic s dng VaR trong
qun tr ri  t bi ng tn tht th tr
nhng tip cn tng t nh c s dng trong nhiu khoa
Hinh.
t trong nh

 ng ri ro th trng cn,
danh mc. S d mm nhng tn tht
v m ca danh ma mn trong danh mc bi
u t c lng m tn thc hi ri ro.

2.3.2 m VaR
VaR ca danh mn th hin m tn th xi vi danh

mn trong mt chu k k n v thi gian) vi m tin cy nhnh.
 2.1: Mu t quyu t mt khon tin lt danh mc c
phia r danh mu t m xung 50000USD.
Sau khi khn nhn st gim li nhun, anh ta mun bit
m tn tht t li ngay lp t mt
ht khon tiu t, nh lp vi thc t t
trng hp thit hi lm khi x ln ti
s ki c bi  n tht t     ng hp s   
4000USD i m ca VaR.
Trong qun tr r s dng r ri ro
m tn tht danh mnh. Cho mt danh m
sung thnh n m ngng sao cho
 tn tht danh mc trong khong thi gian nh
t 

cho trc.

2.3
2.3.3.1 Tip c
Gi s rng mu t quynu t mt danh mn P. Ti thi
 ca danh mu t 
t
V
. Sau mt khong thi gian
t
, ti thi
m
k t t  
 ca danh mu t 
k

V
. 
tk
VVkV  )(
cho
bit s  ca danh mc P trong khong thi gian
t
.
()Vk
g -
 P&L(k))
k
chu k ca danh mc. 

n t:
- u t  v th i vi P sau chu k k nt

P&L(k)
< 0) s b tn tht.
- u t  v th i vi P sau chu k k nt

P&L(k) >
0) s b tn tht.



t
t

tt 


: Biu di n sau khong thi gian
t

 V
k
n ngn ngi F
k

 t c


 x

gPhân vị mức α F
k
. V 


P&L(k) < 0 t   u t trng v s b tn th    x

  
Pr(P&L(k)  x

) = 1 - 

) = 1 -  
u t n v s b tn tht.

V

t


V
k


: Biu din m


u t  v th trng v, khi
0)(  kV
tu t s chu tn
tht. P(
)(kV


 u t chu tn tht di mc x

(x



Ngc lu t  v th n v,
0)(  kV
tu t s chu tn
tht. P(
)(kV
 x


) = 1 - P(
)(kV


) = 1 -  u t chu
mc tn thc x

(x

- 
 hai v th u tu t chu tn tht t
danh mc st gi  hai trng hc cho nh 
mc tn th  
vy VaR ca mt danh mc vi chu k  tin cy (1- 
 F
k
(x). i l 
Nh v
)(kV


 , 

t u t
nm gi danh mc P  sau mt chu k k, v tin cy (1- u t 
 tn tht mt khon s bng
VaR(k, α)
u kin hong.
 2.2: : VaR(1
u USD. Nh vy vt 5%, trong m thng ca

  u USD.
: Trong thc tn quc t:
 Nu chu k c 5%.
 Nu chu k 

2.3.3.2  thit c
 ri ro (VaR) ph thu  mt s
php cn VaR phi tham s 
 ngi quy c  thit ru t
ng    vng b         
tng quan vi nhau. Nc lng mi chui
th thit ca OLS b vi phm. Mt chuc gng nu k
vng, phhip phi theo th
 t ca chui theo thi gian.
 Bc ng : Mt bin
t
Y
    t bc ng  u
ttt
uYY 
1

t
u
u trhng sai
p phng sai b
)()()()(
11 

tttt

YEuEYEYE
.
 vng ca
t
Y
i. Vi gi thii ta tin
r t thu .
 tr  n nht thit ph 
 Thi gian c nh: Gi thit khong thi gian
u khong thi gian. Chng hn, nu cho khong thi gian
mt tu m rng cho m
 i chun: Trong mt s ph thit li su
si chun, ch tr mt s php cn VaR
phi tham s nh Monte Carlo.

2.3


t l







n, 








:
& ( )
& ( )
t t t
t
P L k
r P L k rV
V
  
.
Do
t
V










n, 










a li

t
r
.

2.3i sun
Gi thit chui li sut cn
t
r
i d chun. Vi gi thit
 cn s dng hai tham s k vng (

 lch chun (

) (hoc s
dc lng c  VaR.
T gi thit
),(~
2

Nr
t

suy ra
)1,0(~ N
r
t




nh nh sau:

)(%)100)1(,1(
1
 NngàyVaR
(2.2)

 i chun.
 2.3: u t nm gi mt khi lng c phi 

m hin ti
 V
t
= 100 tri ng, li su    chun r
t

2
( , )N

vi  = 3%.
ng, li su  gi nh
0



. Vi m
 = 5% VaR ca li sut :
-1,64 0,03 = -0,0492.
Suy ra VaR ca danh mc:

t
, 5%) = VaR
Li sut

t
= (- 0,0492)*100 = - 4,92.
Vy sau mt 5%, kh u t  l u.

2.3c
Cho danh mc P: (w
1
, w
2
, , w
N
) vi li su n trong danh mc 
 

n: r
i

2
( , )

i
N

vt :
i
1
w.
N
Pi
i
rr



;
i
1
w.
N
Pi
i
rr



;
2
W'.V.W
P




y li sut ca danh mc r
P

2
( , )
PP
Nr

. T ng t nh i vn
c VaR ca danh mc:
ppr
NngàyVaR
p

)(%)100)1(,1(
1

. (2.3)
: Nc P di d: P: x = (x
1
, x
2

N
) vi x
i
 khon
tiu t n i  

1
& ( )
N
ii
i
P L k rx



.
Vi gi thit li sun trong danh mc r
i

2
( , )
i
N


2
&&
& ( ) ( , )
P L P L
P L k N




2
&&

1
; '
N
P L i i P L
i
x r x Vx




. 
- )100% ) = 
P&L
+ N
-1
()
P&L
= 
P&L
+ N
-1
(x)
1/2
(2.4)
Vi chu k i lng 
P&L
 c t  b 
th 
- )100% ) = N
-1

(
1/2
(2.5)
3: 

 gi nh li sut danh mc (ho
si chu cn c lng hai tham s: k vng ( lch
chun () ( ma trn hip phng sai V v

c)  c VaR, do














ph

. Trong thc t  thit  
  vi phm 

 c l




 . 


 ta 

 ng ph

 i cho c lng VaR gn vi  tn tht
trong thc t nht.

2.3.3.5 n ch c i ro VaR
VaR ca danh mn th hin m tn th xi vi danh
m n trong mt khong thi gian nhnh vi m tin cy nh nh. Tuy
,    VaR 

i 





, VaR mi ch
 mt tn l 
VaR  lt phn nh i (1% hay
ng xu-ng vi nhng din bin bt thng ca th trng), khi xy ra tn
tht, mc tn th d  i ro ES d cho
 li.



2.4 n tht k vng ES (Expected Shortfall)
Nh 

 , mc s dn bin trong qun tr ri ro th
trng, rng ca danh mng hn ch nhnh c 
phng di  t ln thc tin. M  p cn m   ng ri ro th
trng ca danh m   c s dng th  Tổn thất kỳ vọng (Expected
Shortfall  ES). 





 s  u v th  i ro  
phc lng .

2.4.1 m
Sa danh mi nhng trng hp tn
tht thc t ca danh mc vt ng vng) cc
tn thT 



:
Tn tht k vng ca danh mc v tin cy (1- , α), i
lng k vu kin:
( ) ( / ( ))ES ES E X X VaR



  
. (2.6)
Nh mt s t u vit hn VaR, vic s d i ro ES th hin vi
lng r hg VaR.

2.4.2 Mt s t ca ES




ch





sau (xem [10]):
  i ro cht ch ca danh mc.
 M i ro cht ch a danh m biu din nh mt t
hp li ca ES v   (X).
Nh vy via danh mc va thay th VaR 
ng r hn va ch i ro u vit.

2.4.3 Phc nghim c lng ES
 thuc lc
l X ca danh m   i sut (loga li sut) ca danh mc  nh bi:
( 1)
ln
t

t
V
r
V






ng t nh khi c lng VaR t s ling
c lng ES: ph .
 Ph  d  nh v i ca li sut r: chng hn
i chun, T- Student, Pareto t, s li ca r,
s dc lng trong th lng (h
moment t c l c trng ca
 h     c lng ca VaR (xem [9])   ng ng (xem
[10]).
 Ph a ra gi nh v i ca li sut r 
c lng thc nghi
k thup x (phi suy, mng n c lng
(xem [11], [18], [21], [22]).
 c c lng ES nh sau: Cho m      
thng chn  = 1% hoc 5%. Lp mc n: (X
1
, X
2

n
u X

i:n

th t th i ca mu, t
1:n
 X
2:n
 
i:n
 X
n:n
. Gi k n
a n

t p = n

- k. Nu n

 p 
mu c t t n k:
1: 2: :
:

n n k n
kn
X X X
X
k
  

(2.7)

c c lng thc nghit tham kho trong
[10]):
:
()
kn
VaR X


(2.8)

:
: 1:
( : )
()
(1 ) ( : )
kn
k n k n
X n nguyên
ES
p X pX n không nguyên









  



(2.9)
 2.4: Trong  , ta c lng thc nghim ES cho th trng ch
Vi hu a
c thu thp trong khong thi gian t n 6/2010 t ngun
VnDirect. c 

m, ng





 








c (xem [10]): n =1116,

n

= 11,6 
ra k
1

=11, k
2
p
1
=0,6; p
2
=0,8. S dc c lng (2.8), (2.9
c lng thc nghim ci sut th tr
VaR
VnIndex
(1%) = 4,604 (%); VaR
VnIndex
(5%) = 3,686 (%)
ES
VnIndex
(1%) = 4,731 (%); ES
VnIndex
(5%) = 4,249 (%)
T 

 t s nh sau mch t
 Nu li sut th trng gii kh c gi 
i 99% kh 
 ng xu, nu li sut th trng ging 
 c gim d     mc gim d
 
i sut th tri chuy sau khi c
l s dc 2 theo th 
 n. Nu chu k ch) hoc 10


VaR
VnIndex_tun
(1%) = 10,294 (%); VaR
VnIndex_tun
(5%) = 8,242 (%)
ES

(1%) = 14,96 (%); ES

(5%) = 13,43 (%)

2.4.4 
Tn tht k vng ca danh mc v tin cy (1- ).100%, n th


vt ngng
VaR

:
( / )ES E X X VaR


.
Mc tn tht k vng (ES) ca danh mc g xu i ro b
sung cho VaR nhm quan trng cn tr ri r
rDo cc tp h c lng ES c
phc bi cp ti danh mc tp nh 
danh mc ca t chng.

References




Tài liệu tham khảo
1. Acerbi, C., Nordio, C., Sirtori, C.(2001), Expected Sho rtfall as a Tool for
Financial Risk Management, AbaxBank – Working Paper.
2. Danielsson , J.&de Vries, C. ( 1997), Tail index and quantile estimation with
wery high frequency data, Journal of Empirical Finance 4, 241-257.
3. Danniel De Waal (200 4), Statistics of Extremes- Theory and Applications,
John Wiley& Sons, Ltd.
4. Feter F. C hristoffersen (2003), Elements of Financial.Risk.Management.
5. Fotios C. Harmantzis, Linyan Miao, Yifan Chien, Empirical Study of Value-
at-Risk and Expected Shortfall Models with Heavy Tails, Working Paper -
Financial Analytics Group, Stevens Institute of Technology, August 2005.
6. Hull, J. and A. White, Value at Risk When Daily Cha nges in Market Vari-
ables Are Not Normally Distributed, Journal of Derivatives, (1998) 5.
7. Koji Inui, Masaaki Kijima, On the significance of expected shortfall as a
coherent risk measure, Journal of Banking & Finance 29 (2005) 853–864p
8. Longgin. M ( 2000), From value at risk to stress testing: The extreme value
approach, Journal of Banking and Finance 24, pp.1097-1130 .
64
9. Manfred Gilli, Evis Kellezi (2003), An Application of Extremme Value The-
ory for Measuring Risk.
10. McNeil. A. (1998), Calculating Qua ntile Risk Measures for Financial Re-
turn Series using Extreme Value Theory
11. McNeil. A. (1999), Extreme Value Theory for Risk Managers
12. Neftci, S., Value at Risk Calculations, Extreme Events, and Tail Estimation,
The Journal of Derivatives, (2000)
13. R
¨

udiger Frey, Alexander J. McNeil, VaR an d expected shortfall in portfo-
lios of dependent credit risks: Conceptual and practical insights, Journal of
Banking & Finance 26 (2002) 1317–1334p.
14. R D. Reiss & M. Thomas, Statistics Analysis of Extreme Values, with Ap-
plications to Insurance, Finance, Hydrology a nd Other Fields
15. Yasuhiro Yamai, Toshinao Yoshiba, Value-at-risk versus expected shortfall:
A practical perspective, Journal of Banking & Finance 29 (200 5) 997 –
1015p.
65

×